fix(leaks): effects_live_into / effects_live_count and hints_rail_into - a HUD's read-outs into lists it keeps

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

packages/ludic.hints/index.ludic

@@ -6,6 +6,7 @@ import "ludic.base"
 import "defs.ludic"
 import "port.ludic"
 import "rail.ludic"
+import "rail_into.ludic"
 import "card.ludic"
 import "mute.ludic"
 import "system.ludic"

packages/ludic.hints/rail_into.ludic

@@ -0,0 +1,7 @@
+# rail_into.ludic - the rail into a list the caller keeps: a HUD asks every frame, and a list made
+# each time is never given back
+export function hints_rail_into(hints_st: HintsState, out: []int) -> void {
+  List.clear(out)
+  if not HintsWorld.showing() { return }
+  for k in 0 .. hints_st.hn_n { push(out, hints_st.hn_top[k]) }
+}

packages/ludic.hints/tests/hints_test.ludic

@@ -74,6 +74,11 @@ program HintsTest {
     hints_test_st.showing = true
     hints_sort(hints_st)
     expect_eq(len(hints_rail(hints_st)), 1)
+    let kept = new []int
+    hints_rail_into(hints_st, kept)
+    hints_rail_into(hints_st, kept)
+    expect_eq(len(kept), 1)
+    expect_eq(kept[0], hints_rail(hints_st)[0])
   }
 
   test "the key opens the picked chip, the arrows walk round, the key puts it down" (hints_st: mut HintsState, hints_test_st: mut HintsTestState) {